INTRODUCTION — The occurrence of neonatal seizures may be the first, and perhaps the only, clinical sign of a central nervous system disorder in the newborn infant. Seizures may indicate the presence of a potentially treatable etiology and should prompt an immediate evaluation to determine cause and to institute etiology-specific therapy. In addition, seizures themselves may require emergent therapy, since they can adversely affect the infant's homeostasis or they may contribute to further brain injury. Some types of neonatal seizures are associated with a relatively high incidence of early death and, in survivors, a high incidence of neurologic impairment, developmental delay, and postneonatal epilepsy.
Management of neonatal seizures involves accurate diagnosis of seizures, expedited evaluation and targeted treatment for their etiology, and medication to abolish the electrographic seizures. This topic will discuss the approach to treatment of neonatal seizures. The etiology, clinical features and diagnosis of neonatal seizures are discussed separately. (See "Etiology and prognosis of neonatal seizures" and "Overview of neonatal epilepsy syndromes" and "Clinical features, evaluation, and diagnosis of neonatal seizures".)
ETIOLOGIC THERAPY — Treatment directed at the cause of neonatal seizures is critical since it may prevent further brain injury. This is particularly true for seizures associated with some metabolic disturbances (eg, hypoglycemia, hypocalcemia, and hypomagnesemia) and with central nervous system (CNS) or systemic infections. Furthermore, some neonatal seizures may not be effectively controlled with antiseizure medications unless their underlying cause is treated.
The most common etiologies of neonatal seizures are reviewed in the Table (table 1).
Neonatal encephalopathy — Neonatal encephalopathy (and associated hypoxic-ischemic encephalopathy) is the most common cause of neonatal seizures [1]. Even with therapeutic hypothermia for neuroprotection, approximately 50 percent of newborns with hypoxic-ischemic encephalopathy have electrographic seizures [2].
The treatment of neonatal encephalopathy is discussed separately. (See "Clinical features, diagnosis, and treatment of neonatal encephalopathy".)
CNS infection — Neonates with seizures should be presumed to have an infectious etiology until proven otherwise. Thus, a sepsis evaluation is mandatory. Infection of the central nervous system is a relatively common cause of neonatal seizures and should be treated with broad-spectrum antibiotics at doses sufficient to treat meningoencephalitis.
Treatment of infection and meningitis in neonates is discussed separately. (See "The febrile infant (29 to 90 days of age): Outpatient evaluation" and "Bacterial meningitis in the neonate: Treatment and outcome" and "Group B streptococcal infection in neonates and young infants", section on 'Management'.)
Metabolic disturbances — Metabolic disturbances are a treatable common cause of neonatal seizures.
Hypoglycemia — Hypoglycemia should be corrected immediately with a 10 percent glucose solution given intravenously at 2 mL/kg. Maintenance glucose infusion can be given to a maximum of 8 mg/kg per minute. A detailed review of the evaluation and treatment of hypoglycemia in infants is discussed separately. While hypoglycemia is generally a reversible cause of neonatal seizures, severe hypoglycemia can result in brain injury and risk for both refractory neonatal seizures and postneonatal epilepsy. (See "Pathogenesis, screening, and diagnosis of neonatal hypoglycemia".)
Hypocalcemia — Hypocalcemia associated with severe neuromuscular irritability or seizures is treated with 10 percent calcium gluconate (100 mg/kg or 1 mL/kg by intravenous [IV] infusion). The solution is infused over 5 to 10 minutes while the heart rate and infusion site are monitored. The dose can be repeated in 10 minutes if no response occurs. Alternatively, calcium chloride (20 mg/kg or 0.2 mL/kg) can be given. After acute treatment, maintenance calcium gluconate should be added to the intravenous solution. The etiology, evaluation, and treatment of hypocalcemia in neonates are discussed in detail separately. (See "Neonatal hypocalcemia", section on 'Management'.)
Hypomagnesemia — Neonatal hypomagnesemia is often associated with hypocalcemia, although it can occur alone. The treatment is 50 percent solution of magnesium sulfate given by intramuscular injection at 0.25 mL/kg or 125 mg/kg. The same dose can be repeated every 12 hours until normomagnesemia is achieved. (See "Neonatal hypocalcemia", section on 'Correction of hypomagnesemia'.)
Pyridoxine or PLP responsive seizures — Although inborn errors of metabolism are rare, seizures are a common manifestation of many of these disorders, especially in the neonatal period. It is important to recognize such disorders early, since cofactor or vitamin supplementation and other disease-modifying therapies are available for some. (See "Etiology and prognosis of neonatal seizures", section on 'Inborn errors of metabolism'.)
In particular, pyridoxine-dependent epilepsy (PDE) due to antiquitin (ATQ) deficiency and the related disorder, pyridoxamine 5'-phosphate oxidase (PNPO) deficiency, are rare but treatable genetic causes of medically refractory neonatal seizures. The approach to recognition and treatment of PDE is summarized in the algorithm (algorithm 1). Sequential therapeutic trials of pyridoxine (100 mg IV injections, repeated every 5 to 15 minutes up to a maximum of 500 mg with continuous electroencephalography (EEG) monitoring, or 15 to 30 mg/kg per day orally in three divided doses) and pyridoxal 5'-phosphate (PLP, the active form of pyridoxine [vitamin B6]) should be given to neonates with seizures unresponsive to conventional antiseizure medications, particularly if the cause of the seizures is not known.
Trials of IV pyridoxine should be performed with EEG and close cardiopulmonary monitoring, as there is a risk of apnea with pyridoxine, particularly when given IV. If there is no response to pyridoxine or PLP, leucovorin (2.5 mg IV) may be administered, since some cases of antiquitin deficiency respond better to leucovorin (folinic acid) than pyridoxine [3].
The results of one case series caution that EEG-response alone to pyridoxine IV does not definitively identify (nor does lack of initial response exclude) PDE [4,5]. Individuals with pyridoxine- or leucovorin-responsive seizures should undergo further biochemical evaluation including measurement of urine alpha-aminoadipic semialdehyde (alpha-AASA) and/or plasma pipecolic acid [6]. Elevation of alpha-AASA is informative in both treated and untreated states [7,8]. Mutation analysis of the ALDH7A1 gene is recommended in patients with abnormal biochemical screening and/or clear evidence of pyridoxine or leucovorin responsiveness [7,8]. PNPO mutation analysis is suggested in patients with either pyridoxine- or PLP-responsive seizures who have normal alpha-AASA levels.
Patients with antiquitin deficiency should receive chronic supplementation with pyridoxine and/or leucovorin and may also benefit from a lysine-restricted diet supplemented with lysine-free amino acid formula [6,9-11]. Long-term treatment doses of pyridoxine vary between 15 and 30 mg/kg/day for infants [6]. Some commercially available lysine-free formulas are also free of tryptophan, in which case tryptophan should be supplemented. Long-term treatment with high doses of pyridoxine can result in peripheral neuropathy. Infants with PNPO deficiency should receive chronic oral PLP supplementation [6].
Biotinidase deficiency — Biotinidase deficiency due to pathogenic variants in the biotinidase gene may result in medically refractory neonatal seizures that are responsive to oral biotin supplementation. In states where biotinidase enzyme activity is not included in the newborn screening panel, a trial of biotin may be considered in addition to pyridoxine, PLP, and/or leucovorin (algorithm 1).
Molybdenum cofactor deficiency — Molybdenum cofactor deficiency (MoCD) is an autosomal recessive disorder that results from one of several single gene defects in the biosynthetic pathway of molybdenum cofactor [12]. Approximately two-thirds of patients have MoCD type A, in which pathogenic variants in molybdenum cofactor synthesis gene 1 (MOSC1) result in the inability to synthesize the first intermediate in the pathway, cyclic pyranopterin monophosphate (cPMP), and the toxic accumulation of sulfites in blood and urine [13]. Most patients present during the first few days of life with exaggerated startle, lethargy, intractable seizures, and autonomic dysfunction, a complex of symptoms that may resemble hypoxic-ischemic encephalopathy. The disorder can be diagnosed by urine dipstick showing elevated sulfite levels and confirmed with urinary cPMP testing and mutational analysis.
Supplementation of cPMP by daily intravenous infusions is a promising therapy in patients with MoCD type A (but not type B) with the potential to greatly improve neurodevelopmental outcomes when started sufficiently early and continued chronically [14-17]. In February 2021, the FDA approved fosdenopterin, a cPMP substrate replacement therapy, to reduce the risk of mortality for patients with MoCD type A [18]. Approval was based on a combined analysis of several small studies showing an increased survival probability at three years for 13 treated patients compared with 18 matched, untreated patients (84 versus 55 percent). Fosdenopterin is given daily by intravenous injection; the dose is based upon weight and age [19]. The most frequent adverse events include complications related to the intravenous catheter, fever, respiratory infections, vomiting, gastroenteritis, and diarrhea. Because of the potential for photosensitivity, treated patients and caregivers are advised to avoid patient exposure to sunlight and to use protective measures when exposed to the sun.
ANTISEIZURE MEDICATION THERAPY — The use of antiseizure medications for neonates with seizures will be reviewed. Initiating therapy, selecting appropriate agents, and stopping or continuing therapy are the main decisions involved. There are no evidence-based or broadly accepted guidelines for medical management of neonatal seizures, and the approach below is based on clinical experience, observational studies, and a limited number of randomized trials [20-22].
Decision to institute drug therapy — After initial management of airway and cardiovascular support and the identification and institution of etiology-specific therapy, the next decision is whether to initiate antiseizure medication therapy. Factors that must be considered include seizure duration and severity as well as seizure etiology. For example, neonates with brief seizures due to transient, reversible electrolyte or glucose abnormalities do not require immediate treatment with antiseizure medication, while seizures due to other etiologies, particularly if they are prolonged, are properly treated with antiseizure medication.
An historical approach was to treat clinically evident seizures, with or without EEG confirmation of the diagnosis. This approach is problematic because it does not accurately or adequately treat true seizures; infants whose clinical events have no EEG correlate (ie, are not truly seizures) will be exposed unnecessarily to potentially harmful medication, while neonates with clinically subtle or truly subclinical seizures will be insufficiently treated [23]. EEG is therefore critical in the diagnosis and treatment of neonatal seizures. (See "Clinical features, evaluation, and diagnosis of neonatal seizures", section on 'Diagnosis'.)
Medication selection — An approach to first-line and second-line antiseizure medication selection and dosing based on seizure frequency and individual patient characteristics is summarized in the algorithm (algorithm 2). The traditional strategy is to acutely treat seizures with a medication that can be subsequently given as maintenance therapy.
First-line therapy — Phenobarbital has long been used as first-line therapy for seizures in neonates, and it remains the most commonly used agent in this setting [1,24-28]. The next most frequently used first-line agent is fosphenytoin. Enteral absorption of phenytoin is limited for newborns, however, and long-term maintenance dosing of phenytoin is very challenging. Neither agent appears to be more effective than the other and neither is completely effective.
This was demonstrated in a landmark randomized trial of first-line therapy that randomly assigned 59 infants with EEG-confirmed seizures to receive either phenobarbital or phenytoin [29]. Seizures were controlled by first-line therapy in less than half of the infants (43 percent with phenobarbital and 45 percent with phenytoin), and seizure severity was a better predictor of treatment success than the assigned treatment.
Although levetiracetam can be given intravenously, the available evidence does not support its use as a first-line agent for the treatment of neonatal seizures. The NEOLEV2 trial randomly assigned 83 neonates to levetiracetam or phenobarbital as first-line treatment for EEG-confirmed neonatal seizures of any cause [22]. Complete seizure freedom for 24 hours, determined by continuous EEG monitoring, was far more likely for patients assigned to phenobarbital compared with levetiracetam (80 versus 28 percent, absolute risk reduction 52 percent, relative risk 0.35, 95% CI 0.22-0.56]. The greater efficacy of phenobarbital in the NEOLEV2 trial compared with the efficacy of phenobarbital in the earlier trial discussed above [29] was hypothesized to be due to the rapid time to treatment, facilitated by real-time remote review of the neonatal EEG [22].
Thus, existing evidence supports phenobarbital as the first-line agent for neonatal seizure treatment. The initial dose of phenobarbital is typically 20 mg/kg by intravenous [IV] infusion, followed by a maintenance dose of 4 to 6 mg/kg per day in two divided doses. If seizures do not resolve after the first loading dose, repeat boluses of 10 to 20 mg/kg should be given with a goal phenobarbital level of approximately 50 micrograms/mL or a total 24-hour dose of 50 mg/kg (algorithm 2). The use of both phenobarbital and phenytoin in the neonate requires additional knowledge concerning their pharmacologic characteristics [30-34]. (See 'Phenobarbital' below and 'Phenytoin' below.)
Acute treatment can also be initiated with a short-acting benzodiazepine, particularly if a delay is likely prior to availability and administration of a longer-acting medication. (See 'Refractory seizures' below.)
Endpoint of acute therapy — The decision to initiate acute therapy should come with a predefined, expected end point of treatment. Experts advocate the treatment of both clinical and subclinical electrographic seizures, since the only difference between the two types may be their cortical distribution.
Neonates with electroclinical seizures may have electroclinical dissociation, or uncoupling, after treatment initiation. In this scenario, the clinical signs vanish but the electrographic seizures persist [35,36]. Ideal management therefore includes EEG confirmation of treatment response, which is defined most precisely by resolution of electrographic seizures.
The role of continuous EEG monitoring in directing treatment is highlighted in a guideline from the American Clinical Neurophysiology Society [37] and by the International League Against Epilepsy [38]. Since most abnormal movements are not neonatal seizures and most neonatal seizures are subclinical, using EEG to guide treatment of neonatal seizures limits unnecessary exposure to antiseizure medications for those whose events are not seizures and avoids undertreatment of those with subtle or subclinical seizures. However, it is acknowledged that no clinical data prove definitively that this approach improves long-term outcomes.
Current practice consists of urgent administration of antiseizure medication therapy with expedited dose escalation until seizures are controlled, with the first medication given in sufficient doses to achieve seizure-freedom and/or serum levels in the high therapeutic range and/or the maximum tolerated dose. This is followed by additional medications, titrated to effect. (See 'Refractory seizures' below.)
In some cases, seizures cannot be completely controlled with standard treatment and the risks of adverse effects must be weighed against the potential benefit of complete seizure control. The etiology of the seizures is a major factor in this level of decision-making (eg, a target of complete seizure control may be appropriate for a neonate with hypoxic-ischemic encephalopathy but might be unreasonable for a newborn with lissencephaly). The etiology should be reconsidered if the seizures do not respond as expected. As an example, neonatal seizures related to hypoxic-ischemic encephalopathy should resolve within a few days; if not, alternative etiologies such as a metabolic disorder or neonatal-onset epilepsy should be considered.
Refractory seizures — Neonatal seizures refractory to phenobarbital often respond poorly to second-line antiseizure medications. This observation is illustrated by results of a small trial in which neonates whose seizures did not respond to phenobarbital (11 of 22) were randomly assigned to second-line therapy with either clonazepam (n = 3), midazolam (n = 3), or lidocaine (n = 5) [39]. No response was seen in the neonates treated with clonazepam or midazolam. Three of five responded to lidocaine; two neonates became seizure-free with 4 mg/kg per hour of lidocaine, and one had an 80 percent reduction in seizure burden. All of the 11 neonates for whom phenobarbital failed to control seizures had a poor neurodevelopmental outcome at one year.
Importantly, in the NEOLEV2 trial, adding levetiracetam 40 mg/kg did not control seizures for any of the six neonates whose seizures persisted despite two loading doses of phenobarbital, and an additional dose of levetiracetam 20 mg/kg was associated with seizure cessation in only one of the six neonates [22]. Conversely, adding phenobarbital 20 mg/kg when seizures persisted after levetiracetam resulted in seizure control for 14 of 37 infants (38 percent). Thus, levetiracetam loading doses of 40 to 60 mg/kg are unlikely to provide immediate control for as first-line therapy or as second-line treatment for neonates who have persistent seizures after phenobarbital administration.
Choice of a second-line drug in infants who continue to have seizures despite first-line therapy must be individualized, as efficacy data are derived primarily from case series and not from randomized trials. The most commonly used drugs in this setting are phenytoin/fosphenytoin, levetiracetam, lidocaine, and midazolam. Factors to consider when selecting an agent include seizure etiology and severity, the side effect profile of the drug, respiratory and cardiovascular stability of the patient, and the presence of cardiac, renal, or hepatic dysfunction (algorithm 2). Suggested dosing is provided in the algorithm and discussed in more detail below. (See 'Dosing considerations in neonates' below.)
Newer antiseizure medications are increasingly prescribed for neonatal seizures, despite the fact that this is an off-label indication [40,41]. This trend has been driven by incomplete efficacy of more standard agents and concerns about their potential neurotoxicity. However, there is little evidence, and none from randomized controlled trials, that support a greater efficacy and lower adverse event rate with these agents in neonates. The published literature is limited by the lack of standardized medication dosing, variable timing of administration of the newer antiseizure medications, limited EEG monitoring to confirm diagnosis and treatment response, and absence of a placebo arm.
Levetiracetam in particular has been used with increasing frequency, likely due to its readily available intravenous formulation and favorable side effect profile among older children and adults. Practice may change with publication of the NEOLEV2 trial, as the results did not support efficacy of this drug at doses of 40 to 60mg/kg for first- or second-line treatment. None of the infants with persistent seizures after two doses of phenobarbital experienced seizure cessation with subsequent doses of levetiracetam [22]. Levetiracetam may still be a good option in neonates with cardiac or liver dysfunction. Additionally, levetiracetam may reduce neuronal apoptosis in models of neonatal hypoxic-ischemic brain injury [42,43] and might have neuroprotective effects [44]. Despite these encouraging observations, the long-term pharmacokinetic and safety profile of levetiracetam for neonatal seizure treatment are not fully understood and may differ from older children and adults [45-48]. (See 'Levetiracetam' below.)
Intravenous lidocaine is an effective agent for neonatal seizures in selected patients. In cases of continued, EEG-confirmed status epilepticus despite high doses of phenobarbital, lidocaine may be the preferred second-line drug, provided there are no contraindications to its use (eg, congenital heart disease, pretreatment with fosphenytoin/phenytoin) (algorithm 2). In a retrospective study of over 400 full-term (n = 319) and preterm (n = 94) infants with neonatal seizures diagnosed by amplitude-integrated EEG who received lidocaine as a second or third-line agent, the overall response rate was 71 percent [49]. Response rates were higher in full term than preterm infants (76 versus 55 percent). In full-term infants, lidocaine was associated with a higher response rate compared with midazolam in the second-line setting (21 versus 13 percent). Dosing considerations are reviewed below. (See 'Lidocaine' below.)
Continuous infusion of midazolam is also an option in neonates with status epilepticus, provided a secure airway has been established. A nonrandomized retrospective study found that midazolam was rapidly effective in 13 neonates (10 with status epilepticus [SE]) who had electrographic seizures refractory to phenobarbital and phenytoin [50]. Midazolam was given as a bolus of 0.15 mg/kg followed by continuous infusion beginning at 1 mcg/kg per minute and increasing by 0.5 to 1 mcg/kg per minute every two minutes to electrographic seizure control or to a maximum of 18 mcg/kg per minute. Neonates with SE were given a repeat bolus of midazolam 0.10 to 0.15 mg/kg if SE persisted 15 to 30 minutes after the initial bolus. While these results appear promising, randomized clinical trial data are needed to confirm that midazolam is effective for neonatal seizures, especially since midazolam was ineffective in a small, randomized clinical trial [39].
Pyridoxine and pyridoxal-5'-phosphate (PLP) trials should also be considered in neonates with seizures that are refractory to conventional antiseizure medications, particularly if the cause of the seizures is not known (algorithm 1). (See 'Pyridoxine or PLP responsive seizures' above.)
Dosing considerations in neonates
Phenobarbital — Phenobarbital is eliminated by the liver and kidney; thus, infants with impaired hepatic or renal function, such as those with hypoxic-ischemic encephalopathy (HIE), will have a reduced rate of elimination and potential for toxicity with standard dosing. Although therapeutic hypothermia treatment may reduce clearance of phenobarbital marginally, no a priori change in loading or initial maintenance dosing is required [34,51]. The half-life of phenobarbital is greater in premature compared with term infants, and longer in the first month of life compared with older ages in term infants.
Thus, standard phenobarbital dosing in premature infants has the potential for higher serum levels and resultant toxicity. As the infant becomes older, identical daily maintenance doses may result in lower serum levels and create the potential for breakthrough seizures with no other change in the infant's clinical condition. Overall, monitoring trends of serum levels rather than day-to-day fluctuations are more useful in management of phenobarbital therapy [52-54].
A growing body of research on neuronal chloride homeostasis explains, at least in part, why phenobarbital is often incompletely effective in newborns [55,56]. The neuronal chloride gradient in mature neurons is maintained by the activity of potassium-chloride cotransporter 2 (KCC2) channels, which decrease resting intracellular chloride concentrations. When gamma-aminobutyric acid (GABA) receptors are activated (eg, by medications such as phenobarbital), the cell is hyperpolarized due to chloride influx. In immature neurons, KCC2 is underexpressed, whereas sodium-potassium-chloride transporter 1 (NKCC1) channels, which increase intracellular chloride concentrations, are prevalent. The result is a reversed neuronal chloride gradient, such that activation of GABA receptors can paradoxically depolarize the neuron.
These observations have led to interest in bumetanide as a potential adjuvant treatment for neonatal seizures. Bumetanide is a diuretic that acts on NKCC1 channels and could, in theory, be used as rationale polytherapy in combination with phenobarbital. In animal models [57,58] and a human case study [59], co-treatment with bumetanide and phenobarbital appeared to enhance treatment effects. However, an open-label phase I/II trial of bumetanide combined with phenobarbital was closed early due to limited efficacy and important safety concerns, including 3 of 11 surviving infants (27 percent) with significant hearing impairment [60]. Results from a double-blind, randomized controlled trial suggest that bumetanide may be a safe and effective second-line treatment. Seizure burden was reduced, particularly among infants with the highest baseline seizure burden, after combination treatment with phenobarbital and bumetanide. Hearing impairment was observed in 2 of 26 patients (8 percent) treated with bumetanide [61]. The cause of hearing impairment for the five affected infants in these two trials was confounded by the presence of additional risk factors including hypoxic-ischemic encephalopathy in all five and gentamicin treatment in four [60,61], and should be viewed in the context of the approximately 10 percent risk for hearing impairment reported among survivors of hypoxic-ischemic encephalopathy [62]. Nevertheless, until further data are available, routine clinical use of bumetanide as an adjuvant treatment for neonatal seizures is not recommended.
Phenytoin — The prodrug fosphenytoin is the preferred formulation of phenytoin for rapid intravenous loading based on a lower risk of side effects, including a reduced risk of local irritation at the site of infusion. Hypotension and cardiac arrhythmias remain a risk, however, and cardiac monitoring is required. The typical loading dose of fosphenytoin is 20 mg phenytoin equivalents (PE) per kg, at a rate of 3 mg PE/kg/minute.
Pharmacologic characteristics of phenytoin include its nonlinear pharmacokinetics, variable rate of hepatic metabolism, decreased elimination rates during the first weeks of life, and variable bioavailability of the drug with various generic preparations [63,64]. In addition, a redistribution of phenytoin results in a drop in brain concentrations after the first dose. Finally, phenytoin has poor oral bioavailability in infants. Thus, phenytoin use requires individualization of dosing after initiation of therapy and should generally be avoided as a chronic maintenance medication for newborns.
Levetiracetam — The pharmacokinetic and safety profile of levetiracetam for neonatal seizure treatment is not fully understood and may differ from older children and adults [45-48]. It follows that the doses of levetiracetam reported in the literature are very broad (10 to 60 mg/kg/day) [45,46,65]. Based on the NEOLEV2 study, we suggest that levetiracetam should not be administered as a first-line agent [22]. When used in the case of refractory neonatal seizures, we suggest a levetiracetam loading dose of 60 mg/kg IV, followed by a maintenance dose of 60 mg/kg/day IV in three divided doses [47,48].
Lidocaine — Lidocaine is typically administered as an initial bolus dose (2 mg/kg over 10 minutes), followed by a continuous infusion of 7 mg/kg/hour for 4 hours and decreasing the dose by 50 percent every 12 hours for the next 24 hours (ie, 3.5 mg/kg/hour for 12 hours, then 1.75 mg/kg/hour for 12 hours) (table 2) [49]. In order to minimize the risk of iatrogenic arrhythmia, the maximum lidocaine infusion time is 48 hours, but the most recent publications indicate that less than 30 hours is preferable [49,66].
Intravenous lidocaine administration may be arrhythmogenic and requires continuous noninvasive monitoring of electrocardiogram (ECG), heart rate, and blood pressure. Additionally, lidocaine is contraindicated in infants with congenital heart disease and in those who have already received phenytoin/fosphenytoin, due to the heightened risk for arrhythmia [67].
The continuous infusion must be adjusted for neonates treated with therapeutic hypothermia, as hypothermia decreases lidocaine clearance [49]. In this setting, and in infants with low body weight (<2.5 kg), slightly lower doses of lidocaine should be used, although the optimal approach has not been established. Proposed dosing of lidocaine under both normothermic and hypothermic conditions is presented in the table (table 2) [49].
Midazolam — Midazolam is typically given as a bolus of 0.15 mg/kg followed by continuous infusion beginning at 1 mcg/kg per minute and titrated upward to effect [50]. Aside from sedation and the need for assisted ventilation, midazolam is associated with minimal cardiovascular effects.
Duration of therapy — Treatment duration should be guided by the neonatal seizure etiology. As summarized here, most neonates with acute provoked seizures can safely discontinue all antiseizure medications prior to hospital discharge. Conversely, most infants with neonatal-onset epilepsy require long-term treatment.
●Acute provoked seizures – For neonates with acute provoked seizures, we generally discontinue antiseizure medications without a taper after 72 hours of seizure freedom. In a prospective, multicenter study, 303 children with neonatal seizures attributed to an acute provoked cause (ie, hypoxic-ischemic encephalopathy, ischemic stroke, intracranial hemorrhage, or other acute brain injury) were evaluated according to the duration of antiseizure medication treatment, which was maintained at hospital discharge for 64 percent of neonates and discontinued for the remainder [68]. The study was designed and powered to detect noninferiority. At age 24 months, after adjusting for propensity to maintain medications, there was no difference between the antiseizure medication continuation and discontinuation groups for the outcomes of functional neurodevelopment or postneonatal epilepsy development. The results were not different for neonates with severely abnormal EEG patterns or abnormal neurologic examinations. In this study, every infant who developed postneonatal epilepsy before age four months had their antiseizure medication maintained after discharge; thus, maintained medication does not appear to prevent postneonatal epilepsy. These findings support discontinuation of antiseizure medication before hospital discharge for all neonates with acute provoked seizures.
Infants who had ≥3 days of EEG-confirmed neonatal seizures or who have an abnormal neurologic examination at the time of hospital discharge are at increased risk for postneonatal epilepsy. These infants can safely be discharged without antiseizure medication but require close follow-up [69,70]. Parents of infants at high risk should be taught the signs of infantile spasms and other seizure types so they can seek urgent assessment if their child develops postneonatal epilepsy.
●Neonatal epilepsy – In contrast with acute provoked seizures, newborns with neonatal-onset epilepsy syndromes will have ongoing risk for recurrent seizures after the neonatal period and should be maintained on antiseizure medication. Chronic therapy should be tailored to the individual infant. Maintenance doses of phenobarbital are often prescribed (3 to 6 mg/kg per day), and serum levels are monitored. However, for neonates with epilepsies suspected to be caused by channelopathies (eg, pathogenic variants of KCNQ2, KCNQ3 or SCN2a), several reports suggest that carbamazepine or, by extension, oxcarbazepine, may be effective [71-73].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Seizures and epilepsy in children".)
SUMMARY AND RECOMMENDATIONS
●In the neonate, seizures are a medical emergency. They should prompt an immediate evaluation to determine cause and to institute etiology-specific therapy. (See "Clinical features, evaluation, and diagnosis of neonatal seizures", section on 'Etiologic evaluation'.)
●Treatment of the underlying cause of neonatal seizures (for metabolic disorders, central nervous system or systemic infection, or hypoxic-ischemic encephalopathy) is critical since it may prevent further brain injury. Also, neonatal seizures may not be effectively controlled with antiseizure medications unless their underlying cause is treated. (See 'Etiologic therapy' above.)
●Factors that must be considered in deciding upon antiseizure therapy include seizure etiology, seizure duration, and seizure severity. (See 'Decision to institute drug therapy' above.)
●When a decision is made to initiate antiseizure medication, we suggest first-line treatment with phenobarbital rather than phenytoin (Grade 2C). Phenobarbital and phenytoin were equally effective in a randomized trial, but maintenance oral dosing of phenytoin in the newborn is very challenging. In another randomized trial, phenobarbital was clearly more effective than levetiracetam for short-term seizure control as a first- and second-line agent. Dosing schedules are listed in the figure (algorithm 2). (See 'Medication selection' above and 'First-line therapy' above.)
●Neonatal seizures refractory to phenobarbital often respond poorly to second-line antiseizure medications. The most commonly used drugs in this setting are phenytoin/fosphenytoin, levetiracetam, lidocaine, and midazolam. Factors to consider when selecting an agent include seizure severity, the side effect profile of the drug, respiratory and cardiovascular stability of the patient, and the presence of cardiac, renal, or hepatic dysfunction (algorithm 2). (See 'Refractory seizures' above.)
●Pyridoxine (100 mg by intravenous infusion in repeated doses with continuous electroencephalography [EEG] monitoring, or 15 to 30 mg/kg per day orally in three divided doses) and pyridoxal 5'-phosphate (PLP, 60 mg/kg per day orally in three divided doses) should be given sequentially to neonates with seizures unresponsive to conventional antiseizure medications, particularly if the cause of the seizures is unknown (algorithm 1). If there is no response to pyridoxine or PLP, leucovorin (2.5 mg intravenously) may be administered for possible leucovorin (folinic acid) responsive seizures. (See 'Pyridoxine or PLP responsive seizures' above.)
●Current best practice consists of continuing acute medication therapy until all seizures (clinical and EEG seizures) are controlled, with the first medication given in doses sufficient to achieve serum levels in the high therapeutic range or to the maximum tolerated dose before additional medications are added, unless the risks of treatment outweigh the potential benefit. (See 'Endpoint of acute therapy' above.)
●For neonates with acute provoked seizures, we generally discontinue antiseizure medications without a taper after 72 hours of seizure freedom and prior to hospital discharge, even for neonates who have abnormal EEG patterns or an abnormal neurologic examination. Infants at high risk for postneonatal epilepsy should have careful follow-up with a pediatric neurologist. By contrast, newborns with neonatal-onset epilepsy syndromes will have ongoing risk for recurrent seizures after the neonatal period and should be maintained on antiseizure medication. (See 'Duration of therapy' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Eli M Mizrahi, MD, who contributed to an earlier version of this topic review.
